7/29/2017

1

Diagnostic Ultrasound Imaging Quality

Assurance

James A. Zagzebski1, Ph.D.

Zheng Feng Lu2, Ph.D. 1Dept. of Medical Physics University of Wisconsin, Madison

2Dept. Of Radiology, University of Chicago, Chicago

Purpose

• Outline a QA program that is

Responsive to clinical US lab accrediting bodies, ACR and AIUM

Effective at detecting some important system flaws

Can be carried out effectively by medical physicists

• Discuss advanced tools that may enhance or even serve

as an alternative to methods that will be discussed

UltraIQ analysis software for phantom images

Aureon transducer tester

• Introduce Doppler tests (currently not required by ACR)

Information on US QA • Goodsitt M M et al 1998 Real-time B-mode ultrasound quality control test

procedures. Report of AAPM Ultrasound Task Group No. 1 Med. Phys. 25 1385

• IEC 61391-1 (2006) Ultrasonics – Pulse-echo scanners – Part 1: Techniques for

calibrating spatial measurement systems and measurement system psf response

• IEC 61391-2 (2010) Ultrasonics– Pulse-echo scanners – Part 2: Measurement of

maximum depth of penetration and local dynamic range (1996)

• IEC 62736 Ultrasonics (2016) – Pulse-echo scanners – Simple methods for

periodic testing to verify stability of an imaging system’s elementary

performance

• AIUM 2014, AIUM Quality Assurance Manual for Gray Scale US Scanners.

• King et al, Evaluation of a low cost liquid ultrasound test object for detection of

transducer artefacts. Phys. Med. Biol. 55 (2010) N557-570.

• Hangiandreou NJ et al, Four-year experience with a clinical ultrasound

quality control program. Ultrasound in Med. & Biol. 37: 1350-57, 2011.

Information From US Accreditation Bodies

• Ultrasound Accreditation Program Requirements, Am

College of Radiology, (3/22/17 rev) http://www.acraccreditation.org/~/media/ACRAccreditation/Documents/Ultrasound/Requir

ements.pdf?la=en

• ACR-AAPM Technical Standard for Diagnostic Medical

Physics Performance Monitoring of Real Time Ultrasound

Equipment. (2016) https://www.acr.org/~/media/ACR/Documents/PGTS/standards/US_Equipment.pdf

• AIUM 1998, American Institute of Ultrasound in Medicine,

Routine Quality Assurance for Diagnostic Ultrasound

Equipment. http://aium.s3.amazonaws.com/resourceLibrary/rqa.pdf

Annual Surveys, Routine QA (ACR) • Acceptance testing, 6-month Routine QC: optional

• Annual surveys: required

Physical and and mechanical inspection; sterility

Image display performance

Image Uniformity

• Element “dropout” and other sources on non-uniformity

System sensitivity and/or penetration capability

Geometric measurement accuracy during program initiation (optional for annual survey)

Contrast resolution, spatial resolution: optional items for annual survey. http://www.acraccreditation.org/Modalities/Ultrasound

Physical and Mechanical Inspection, ACR

Console

Air filters Lights, indicators Wheels, wheel locks Proper cleaning (are procedures in place?) Viewing monitor, keyboard clean Other safety issues

Air filters

Before After

7/29/2017

2

Image Display (Scanner and PACS)

• Important for monitor on machine to be set up properly to view all echo levels available and entire gray bar pattern. Set up during acceptance testing

Take steps to avoid casual adjustments (mark or inscribe contrast and brightness controls)

• Most machines provide one or more gray scale test patterns for setup and for routine QC. are all gray bars visible? (System,

PACS)

www.philips.com

Gray bar on GE Logiq 9

Image Display (Scanner and PACS)

• Gain and sensitivity adjustments done using system monitor

• Intrepretation most often done on a PACS workstation.

• Important that there is agreement between image features viewable on PACS and the features seen on the system monitor.

• We were finding that the 15 gray bar pattern built into the machine was not sensitive enough to subtle, but important faults in monitor agreement.

10

SMPTE, TG18 or Other Gray Scale Test Pattern

• Available on most

scanners

• 0% to 100% gray

bar pattern

• Squares for

detecting geometric

distortion

• Are all gray

transitions visible?

• Is the 0-5%

transition visible?

• Is the 95-100%

transition visible?

TG18: Q=0+14 l=1

Q=128+14 l=129

Q=255-14 l=254

Monitor agreement (cont.) System Worksheet, page 2 of Report for each scanner

General Machine Cleanliness: Keyboard and knobs clean? ☒Yes ☐No Monitors Clean? ☒Yes ☐No Air Filters clean? ☐Yes ☒No

Mechanical and Electrical:

Wheels fastened securely and rotate easily? ☒Yes ☐No Wheel locks work well? ☒Yes ☐No Accessories fixed securely? ☒Yes ☐No Cords attached securely? ☒Yes ☐No

PACS Workstation-System Monitor Contrast and Brightness between scanner and workstation: ☐1 poor ☐2 ☐3 average ☒4 ☐5 excellent

Assessment made from Both 1 & 2 below: Generate a gray bar pattern. Save it to PACS.

Number of gray levels seen on the system monitor 15+

Number of gray levels seen on the PACS 15+ *Gray bar visualization: With “patient” registered, push “exam utilities;” push “test pattern.” Record an image and compare to the workstation

Count the number of gray levels seen in the room and on the PACS monitor. SMPTE Pattern: 0-5% transition: seen on system monitor: NO seen on PACS: YES 95-100% transition: seen on system monitor: YES seen on PACS: YES

For low level echo detectability, do probes ‘Depth of Penetration’ results judged on the system monitor agree what you would have chosen if judging on PACS? ☒Yes

Generate a gray bar pattern. Save it to PACS.

Number of gray levels seen on the system monitor 15+ Number of gray levels

seen on the PACS 15+ SMPTE Pattern: 0-5% transition: system monitor: NO PACS monitor: YES SMPTE Pattern: 95-100% transition: system monitor: YES PACS monitor: YES

Routine QA: Transducers Check all transducers on the system

(most facilities have many interchangeable probes that float among systems; a systematic approach to evaluate all probes should be in place.

Transducer Inspection Delaminations Frayed cables Proper cleaning

www.providian.com

Tests using phantoms. Current materials:

Water-based gels

Advantages: Speed of sound = 1540 m/s

Attenuation ~ proportional to frequency (specific attenuation expressed as 0.5 or 0.7 dB/cm-MHz)

Backscatter

Disadvantages:

Subject to desiccation (?)

Must be kept in containers

Requires scanning window

7/29/2017

3

• Solid, non-water-based materials

(urethane)

• Advantages: Not subject to desiccation

No need for scanning window; possibility for soft, deformable scanning window

Produce tissue-like backscatter

Disadvantages: C= 1430-1450 m/s

Attenuation ~ proportional to f1.6

Surface easily damaged if not cleaned regularly to remove gels

Tests using phantoms. Current materials: Phantom test 1: Image Uniformity

- Done with each transducer - This example is not a transducer fault, but a TGC problem

20

Image Uniformity

Considered to be the

most important and

useful test!

Ideally: No loss of sensitivity near

edges of the image

No discontinuities between

tx focal zones

No evidence of element

dropout

No vertical ‘shadows’

20

Non-Uniformity caused by element dropout

Most frequent fault

seen in QA testing

Image a phantom

using good coupling

Search for

“shadows”

emanating from the

transducer

Common in new and

old probes!

Disable spatial compounding cross-beam Sono-CT Sea Clear Use single, shallow transmit focus

Need Proper Technique to Detect Element Dropout

Transducer with moderate element dropout Spatial compounding disabled

Difficulties with Uniformity

• Visualizing 1-2 element dropouts

• Use persistence; translate transducer.

7/29/2017

4

Objective Criteria being developed • IEC 62736 Ultrasonics (2016) – Pulse-echo scanners – Simple methods for

periodic testing to verify stability of an imaging system’s elementary

performance

• AAPM Ultrasound Subcommittee Task Group

• Record a cine loop while translating the transducer to the image plane.

• Compute the ‘median’ image for this (~100) image loop

• Plot a lateral intensity profile from a ~3-10 mm axial range

• A dip >3dB and more than 2 elements wide is worth counting as a defect of

possible concern.

Median image

Dip magnitude and width analyzed in uniformity assessment

Median Image

Image Uniformity(Automated QC Software)

Difficulties with Uniformity (coupling)

• Solution 1: rock transducer from side to side

30

Difficulties with Uniformity (coupling)

• Solution 2: Use a liquid or easily deformable TM material

Electronic Probe test King et al, Evaluation of a low-cost liquid ultrasound test object for detection of transducer artifacts. Phys. Med. Biol. 55 (2010) N557-570.

Liquid Conventional

Difficulties with Uniformity (coupling)

• Solution 3: Use a phantom having concave windows

(Goodsitt et al, AAPM Ultrasound Task Group work)

(AIUM 2014, AIUM Quality Assurance Manual for Gray Scale Ultrasound Scanners, manufactured by Ernest Madsen, Univ. of Wisc.)

Difficulties with Uniformity (coupling)

• Solution 3: Use a phantom having concave windows

Gammex 410

7/29/2017

5

Transducer worksheet part of UW Report

Transducer

ID/Serial

Number

Cables/

cracks/

delaminate

Uniformity,

dropout

Sensitivity (Depth of

Penetration)

(MHz/cm)

Geometric Accuracy

H: cm/actual cm

V: cm/actual cm Conclusions and

recommendations

OK No OK No

C1-5

79635YP9 ☒ ☐ ☒ ☐

5MHz/13.71cm

H5MHz/10.6 cm

H: 5.81/6

V: 8.01/8

Uniformity Rating 1

DOP ≈ to previous results

☒ Yes ☐ No

Click here to enter

comments.

Instructions, uniformity ratings (UW-Madison, not other groups, such as AAPM): 1=uniform 2=minor inhomogeneity (no more than 2 minor dips) 3=Significant inhomogeneities; transducer is functional, but consider replacing 4=Immediate repair or replacement recommended Data table (1 line for each transducer)

Device to test ultrasound transducers • 2D matrix receiver captures energy

profile of transducer while running on the scanner system

• All 1-D and 2-D transducers from any manufacturer

• All operating modes, including ARFI and shear wave imaging

• Assesses lens stability over time • Potential to calculate acoustic dose

Aureon by ACERTARA http://www.acertaralabs.com

Sensitivity, Maximum Depth of Penetration

• Considered by many as a good overall check of the integrity of the system

• FOV at 18 cm (or set to match the phantom/transducer capabilities)

• Output power (MI) at max

• Transmit focus at deepest settings

• Gains, TGC for visualization to the maximum distance possible

Maximum “Relative” Depth of Penetration

How far can you see the speckle pattern in the material?

40

Objective Maximum Depth of Visualization • Shi, Al-Sadah, Mackie, Zagzebski, Signal to Noise Ratio Estimates on

Ultrasound Depth of Penetration (abstract only), Medical Physics 30: 11367,

2003.

• Gorny, Tradup, Bernatz, Stekel, and Hangiandreou, “Evaluation of an

Automated Depth of Penetration Measurement for the Purpose of Ultrasonic

Scanner Comparison”, (abstract only), J. Ultrasound Med 23: S76, 2004.

• Rubert, et al, Automated Depth of Penetration Measurements for Quality

Assurance in Ultrasound (Abstract only), Medical Physics 342, 11367, 2015.

• Specified in IEC International Standards 61391-2 (2010) and 62736

(“Maximum Relative Depth of Penetration” in 62736)

• Compute mean pixel value vs. depth for

phantom (signal+noise) and for noise only (noise)

• Depth where (signal+noise)/noise = 1.4 =DOP

MPV’

IEC Standard 61391-2: Automated Method

3. Average ROI Data Horizontally

in Both Images

2. Record an in-air “noise”

image

1. Record a phantom image

7/29/2017

6

DOP testing: manual and automated

Phantom In-Air

1.4 × noise in air

Signal + noise in phantom

Depth (mm)

Avera

ge P

ixel Valu

e

Transducer

ID/Serial

Number

Cables/

cracks/

delaminate

Uniformity,

dropout

Sensitivity (Depth of

Penetration)

(MHz/cm)

Geometric Accuracy

H: cm/actual cm

V: cm/actual cm Conclusions and

recommendations

OK No OK No

C1-5

79635YP9 ☒ ☐ ☒ ☐

5MHz/13.71cm

H5MHz/10.6 cm

H: 5.81/6

V: 8.01/8

Uniformity Rating 1

DOP ≈ to previous results

☒ Yes ☐ No

Click here to enter

comments.

Instructions, uniformity ratings (UW-Madison, not other groups, such as AAPM): 1=uniform 2=minor inhomogeneity (no more than 2 minor dips) 3=Significant inhomogeneities; transducer is functional, but consider replacing 4=Immediate repair or replacement recommended Data table (1 line for each transducer)

UW Report Transducer worksheet (page 3)

Distance Measurement Accuracy: Vertical

Actual 8.0 cm

Measure 7.94 cm

error 0.75%

Acceptable

*Action: >1.5mm or 1.5%

*Defect: >2mm 0r 2%

*Goodsitt M M et al 1998 Real-time B-mode ultrasound quality control test procedures. Report of AAPM Ultrasound Task Group No. 1 Med. Phys. 25 1385

Distance Measurement Accuracy: Horizontal

Actual 6.0 cm

Measure 6.05 cm

error < .8%

Acceptable

*Action: >2mm or 2%

*Defect: >3mm 0r 3%

*Goodsitt M M et al 1998 Real-time B-mode ultrasound quality control test procedures. Report of AAPM Ultrasound Task Group No. 1 Med. Phys. 25 1385

Routine QA (ACR General US Program) • Distance Measurement

Accuracy tests

Necessary? (“Scanner is a

transducer tied to a computer.”)

May be important for specific

uses

• Images registered from 3-D

data sets

• Workstation measurements

• Radiation seed implants

Reconstructed Elevational Plane Acquisition Plane (Normal 2-D view)

49

Routine QA (ACR General US Program) • Distance Measurement

Accuracy tests

• Required in the

mechanically scanned

direction

3-D 2-D

Scan plane is perpendicular to previous views

7/29/2017

7

Routine QA (ACR General US Program)

Reconstructed Elevational Plane Acquisition Plane (Normal 2-D view)

3-D 2-D

• Distance Measurement

Accuracy tests

• Required in the

mechanically scanned

direction

Actual: 6.0 cm Measured: 6.04 cm Error: <0.7%

UW Report Transducer worksheet (page 3)

Transducer

ID/Serial

Number

Cables/

cracks/

delaminate

Uniformity,

dropout

Sensitivity (Depth of

Penetration)

(MHz/cm)

Geometric Accuracy

H: cm/actual cm

V: cm/actual cm Conclusions and

recommendations

OK No OK No

C1-5

79635YP9 ☒ ☐ ☒ ☐

5MHz/13.71cm

H5MHz/10.6 cm

H: 5.91/6

V: 8.01/8

Uniformity Rating 1

DOP ≈ to previous results

☒ Yes ☐ No

Click here to enter

comments.

Instructions, uniformity ratings (UW-Madison, not other groups, such as AAPM): 1=uniform 2=minor inhomogeneity (no more than 2 minor dips) 3=Significant inhomogeneities; transducer is functional, but consider replacing 4=Immediate repair or replacement recommended Data table (1 line for each transducer)

Spatial Resolution? • Not done routinely

2 image sets, each taken with a

different speed of sound

assumption in the beam former

Targets not agreed on

universally

• Anechoic objects get fuzzy

with poorer resolution

• Line targets get wider

Requires standardized gain

settings to make meaningful

Enhance using computational

methods to measure point

spread function width?

Image of a phantom is useful for qualitative

comparisons!

Images obtained during routine Breast QC testing, 3/2010

Conventional Spatial Compounding

Equipment Evaluation Tests Pass/Fail Comments

1. Physical and

Mechanical Inspection Pass

2. Image Uniformity

and Artifact Survey

6 Probes Pass

0 Probes Fail

There is an obvious (small) region of element

dropout in the C1-6 transducer. Other transducers exhibit no apparent dead elements.

3. Geometric Accuracy

6 Probes Pass

0 Probes Fail

4. System Sensitivity

6 Probes Pass

0 Probes Fail

5. Scanner Electronic

Image Display Performance

Pass The SMPTE test pattern 0%-5% transition is not

seen on the display, but is seen on PACS images.

6. Primary

Interpretation Display Performance*

Select one. All gray level transitions in video test pattern seen; all

transitions on SMPTE pattern also seen.

7. Contrast Resolution

(Optional)

Enter #. Probes

Pass

Enter #. Probes Fail

☒ Not Tested

8. Spatial Resolution

(Optional)

6 Probes Pass

0 Probes Fail

☐ Not Tested

Images of resolution test zones of the phantom are

obtained for reference.

Medical Physicist's (or designee's) Recommendations for Quality Improvement:

The purpose of this program is to: inspect mechanical features and cleanliness of the imaging

system; evaluate adequacies of image monitor settings; inspect transducers and check for

flaws, dead elements, loss of sensitivity; and assess geometric accuracy (calipers); and record

an image for use in assessing consistency of resolution for each probe. During acceptance

tests, Doppler also is assessed.

The system is operating well with all probes. The SMPTE test pattern offers an additional,

challenging low level gray transition (0% - 5 %) which often is not easily visualized on the system monitors, and this is the case for this machine.

Generally, the system is operating well.

Site, Location, Facility UAP* number, Date, Physicist

Machine ID, PACS ID

UW Report (page 1, Summary)

*ACR Ultrasound Accreditation Program

4-year Experience with a clinical ultrasound quality

control program, (Hangiandreou et al., Ultrasound Med Biol 37, 1350-1357, 2011)

Evaluation Method # of detected

“failures”

% of detected

“failures”

Mechanical Integrity 47 25.1

Image uniformity 124 66.3

Distance Accuracy 0 0.0

DOP (penetration) 3 1.6

Clinical Problems 13 7.0

TOTAL 187 100.

Recommendation

Quarterly

Quarterly

Annually

Annually, (if done with

software)

Sonographer’s daily

inspections

7/29/2017

8

Future

• Incorporate computational methods for more objective tests

• Expand to other operating modes Pulsed Doppler

• Sensitivity (signal to noise at a given depth, for both fast and slow flow conditions)

• Velocity accuracy

• Directional Discrimination; gate accuracy, etc.

• Volume flow

• QIBA volume flow project (just starting)

Color flow

Elasticity, shear wave (SW) imaging

• QIBA work on SW velocity in liver (advanced stages)

Doppler Effect in Medical Ultrasound

time

Gammex 1425 Flow Phantom Doppler 403 Flow Phantom

33 cm/s

32 CM/S 33 CM/S

33 cm/s Volume Flow rate = 10 ml/s Peak velocity 101.8 cm/s

With a wedge offset, tilting the transducer to enable a 60o Doppler angle.

Volume Flow rate = 5 ml/s Peak velocity 50.9 cm/s

Doppler Effect in Medical Ultrasound

time

Gammex 1425 Flow Phantom Doppler 403 Flow Phantom, Siemens S2000

33 cm/s

32 CM/S 33 CM/S

33 cm/s

0

20

40

60

80

100

120

140

0 20 40 60 80 100 120 140

Vel

fro

m D

op

ple

r (c

m/s

)

Theoretical Vmax (cm/s)

Velocity from Spectral Display vs. Theoretical Velocity

Series5Series6Series7Series8Series1

Diagonal segment Horizontal segment, wedge Expected (theory) .

Doppler Effect in Medical Ultrasound

Medical Physics Dept.

time

Gammex 1425 Flow Phantom Directional Accuracy, Doppler

33 cm/s

32 CM/S 33 CM/S

33 cm/s

System with poor directional detection. Flow appears to be bidirectional, even though it is only from right-to left..

Continuous flow Pulsed flow Distance along array (%)

Doppler Effect in Medical Ultrasound

Medical Physics Dept.

time

Gammex 1425 Flow Phantom Effects of Dead Elements in Transducers

33 cm/s

32 CM/S 33 CM/S

33 cm/s

Doppler beam line emerging from area of array exhibiting element dropout. (Weaker spectral signal, lower apparent velocity.)

Doppler beam line emerging from array region with no evidence of element dropout. It was recommended probe be replaced

Summary

• Setting up, maintaining an equipment QA program is straight forward

• The ACR listed procedures form a useful, basic QA program

Directed by physicist or lab personnel

Integrated effort including lab and technical staff

Requires a Phantom

Closely correlates with AIUM list of factors needing to be tested

• Transducer uniformity problems, element dropout, a frequent fault in today’s scanning machines

• Computational methods can be developed for objective tests